Main

PWS is the most common syndromal cause of human obesity. It is a complex multisystem disorder caused by a paternal deletion or maternal disomy of chromosome 15q11-13(1), and is characterized by perinatal hypotonia with sucking and feeding difficulties, hypogonadism, later hyperphagia leading to obesity, short final stature, psychomotor developmental delay, behavioral abnomality, and dysmorphic features(2). Obesity is the rule in patients with uncontrolled PWS, because their caloric intake is high and their energy expenditure is low(3). It has been suggested that the hypothalamic dysfunction observed in these subjects may lead to lack of satiety, and thereby an increased appetite.

Satiety is one of the most important factors controlling energy intake. The identification of the ob gene(4) and its corresponding product leptin as well as the leptin receptor(5) has resulted in a large number of reports that have demonstrated the great pathophysiologic importance of the leptin signaling pathway for both eating behavior and energy balance in rodents, whereas the physiologic role of this system in man is less well known(6). The ob gene is highly expressed in fat cells from which circulating leptin is secreted. This protein is presumed to mediate signals from adipose tissue to the brain, thereby acting as a lipostat mechanism through modulation of satiety signals(7). If no leptin is produced, satiety is not obtained, and the subject consequently continues to eat. In obese humans, however, increased gene expression and elevated circulating levels of apparently normal leptin have been observed(8–13). The latter might indicate that central resistance to leptin is of importance for the development and/or maintenance of human obesity. For this reason, PWS with its hypothalamic dysfunction is a particularly interesting disorder for studies of the role of the leptin system in man. We have measured the human ob gene expression in s.c. adipose tissue and the corresponding serum leptin levels in children with PWS, children with nonsyndromal obesity, and nonobese children to reveal possible mechanisms explaining differences in the body weight regulation between these groups.

METHODS

Subjects. Six prepubertal children (one girl and five boys, aged 4.3-12.7 y, BMI 0.3-6.1 SDS) with clinically and molecular-genetically verified PWS and more than 8 points of the internationally accepted PWS scoring system(14), six nonsyndromal obese children (one girl and five boys, aged 10-16 y, BMI 3.2-8.6 SDS), and 20 healthy prepubertal nonobese children (five girls and 15 boys, aged 0.2-12 y, BMI -1.9 to 1.8 SDS) were included for mRNA measurement. Fasting venous blood samples for determination of serum leptin were obtained from the obese children, the children with PWS, and 23 healthy nonobese prepubertal children (three girls and 20 boys, aged 5.5-14.6 y, BMI <2 SDS)). No significant difference in age was found between the obese (PWS and nonsyndromal) children and the nonobese children (9.7 y ± 3.7 versus 11.4 y ± 3.3). Except for PWS, obesity, or hernia inguinale, the children had no apparent disease and took no medication.

Adipose tissue biopsies. Children with PWS and nonsyndromal obesity had abdominal s.c. biopsies taken after local anesthesia with prilocaine. Adipose tissue from nonobese children was obtained from elective inguinal hernia surgery. The children had been fasting overnight, and the biopsies were taken in the morning. The adipose tissue specimens, which ranged from 0.2 to 0.5 g, were immediately frozen unfixed on dry ice or liquid nitrogen and stored at -70 °C until sectioned.

The study was approved by the Ethics Committee of the Karolinska Institute, Stockholm, Sweden. All parents gave informed consent to participation in the study.

In situ hybridization. One oligonucleotide complementary to nucleotides 122-169 (ob 1) of the human ob sequence(4) and one γ-actin oligonucleotide, complementary to nucleotide 325-372, were synthesized, purified, and labeled as previously described(9). The γ-actin probe was used to secure, first, that the mRNA was of equally good quality in all examined tissue samples and second, that changes in leptin mRNA were specific for the latter gene.

Adipose tissue samples from all 32 human subjects were analyzed with respect to ob and γ-actin expression in s.c. fat. Adipose tissue was cut at 20-μm thickness in a cryostat, thawed onto electrically charged Fisher probe on + slides (Fisher Biotech), and processed for in situ hybridization as previously described(15).

Autoradiographic(14) C microscale strips (Amersham Corp.) ranging from 2 to 98 nCi/g were used as standards by coexposure with the tissue sections on Amersham Hyperfilm, B-max, X-ray film. Samples from three PWS, three other obese, and three nonobese subjects were mounted on each slide to simplify the comparison between the different groups and between different individuals. In all subsequent steps, slides were processed together, effectively handled as a single batch through hybridization, washes, autographic exposure, and computerized image analysis. Three sections from each individual tissue were included in the analysis. The computer image was performed as previously described(9).

Leptin assay. The circulating plasma leptin levels were determined with a commercially available human leptin RIA kit (Linco Research Inc., St. Charles, MO). The antibody was raised against highly purified human leptin, and both the standard and tracer were prepared with human leptin. The samples were run in duplicate. All samples were in the detection range of the kit, i.e. 0.5-100 ng of leptin/mL, and the intra-and interassay coefficients of variance were 3.9 and 4.7%, respectively.

Statistical analysis. Values are given as the mean ± SD and median ± range. The two-tailed unpaired t test, ANOVA, and single regression analysis were used for statistical comparisons.

RESULTS

BMI in nonsyndromal obese children was significantly increased compared with children with PWS (5.2 SDS ± 1.9 versus 2.5 SDS ± 2.1, p = 0.05). BMI of children with PWS and nonsyndromal obesity was significantly increased compared with nonobese children (3.8 SDS ± 2.4 versus -0.8 SDS ± 1.0, p ≤ 0.001).

The individual leptin mRNA values in s.c. adipose tissue from children with PWS, nonsyndromal obesity, and nonobese children are depicted in Figure 1A. To adjust for minor differences in general mRNA content in different preparations, leptin mRNA is expressed as the ratio between leptin mRNA and γ-actin mRNA, which has been shown to be expressed to the same extent independently of the BMI(10). The leptin mRNA expression ratio in adipose tissue from PWS and nonsyndromal obese children was in a similar range (103-246% and 106-171%, respectively) and was significantly increased compared with the leptin mRNA expression ratio in adipose tissue from nonobese children. Leptin mRNA levels in nonobese children varied considerbly (25-94%). Four of the 20 nonobese children had an elevated leptin mRNA expression ratio. These four children were the oldest ones in the nonobese group (5-12 y of age). We have investigated the parental BMI of all nonobese children. In three of these four children the parental BMI of one or both parents was >27 kg/m2, whereas the parental BMI of the others was in the normal range.

Figure 1
figure 1

(A) Individual leptin mRNA expressed in children with PWS, nonsyndromal obese children, and nonobese children. The values of each subject is the mean value of three adipose tissue sections processed in one in situ hybridization experiment. The median and the range are expressed for the different groups of children. There was a significant difference between children with PWS/nonsyndromal obesity and nonobese children (p < 0.01). (B) Individual serum leptin levels of children with PWS, children with nonsyndromal obesity, and nonobese children were determined by a commercially available kit. The different group values are median and range. The levels of serum leptin (S-Leptin) were significantly higher in children with PWS and nonsyndromal obesity than in nonobese children (p < 0.001). (C) Simple linear regression analysis showing the correlation between serum leptin levels and BMI SDS in the total material of children with PWS and nonsyndromal obesity and in nonobese children (rs = 0.80, p < 0.005).

Full size image

Fasting serum leptin levels in children with PWS and nonsyndromal obesity were not significantly different (Fig. 1B). However, compared with nonobese controls, the serum leptin levels in both children with PWS and nonsyndromal obesity were significantly higher (32.6 ng/mL ± 20.8 versus 4.9 ng/mL ± 3.1, p < 0.001). As depicted in Figure 1C, serum leptin also correlates to BMI SDS (r = 0.80 and p < 0.005) as previously shown in childhood studies on leptin(12, 13). In the present study a correlation was found between leptin mRNA and BMI (r = 0.73 and p < 0.001). Furthermore, leptin mRNA was found to correlate to age at least in nonobese children (r = 0.62 and p = 0.004). No difference in leptin mRNA expression or serum leptin levels was found between genders.

DISCUSSION

This study confirms that obesity is associated with increased leptin mRNA and serum leptin levels(9–13). Furthermore, serum leptin levels were correlated to BMI when both lean and obese children were included. On the other hand, no significant correlations were observed between serum leptin levels and BMI or between serum leptin and leptin mRNA when the obese children were studied separately. This is in contrast to previous reports in adult subjects(12) and may be due to the size of the material. However, differences in leptin regulation between children and adults cannot be completely excluded; therefore, it is possible that circulating leptin to a lesser extent is regulated at the transcriptional level in children. From the present data we cannot conclude whether the higher ob gene expression in the obese children is due to a higher transcriptional rate or to a slower mRNA turnover. It still remains to be demonstrated whether leptin in man is involved in the regulation of satiety and, if so, to what extent total plasma levels or leptin levels in the cerebrospinal fluid are of importance. At least in adult obesity, data have been reported that strongly suggest a blunted transport of leptin over the blood-brain barrier(16, 17). From previous knowledge about PWS it is tempting to speculate that this disorder would be due to a specific defect in the hypothalamic satiety center and that the mechanism underlying obesity would be different from that in nonsyndromal obesity. However, in the present study no significant differences in leptin mRNA and serum leptin levels were observed between these two forms of obesity. From the present data, it seems likely that neither PWS nor nonsyndromal childhood obesity is caused by a defective regulation of fat cell leptin expression. The feedback response to a putative leptin receptor resistance seems to be similar in these two disorders, but the mechanisms behind the resistance may of course be different in PWS compared with nonsyndromal obesity. The hypothalamic mechanism behind the weight-regulating effects of leptin are yet only partly revealed. However, several reports indicate that neuropeptide Y may have a major role in mediating leptin signals(6, 7, 18).

Nevertheless our mRNA and serum leptin data are compatible with an overstimulation of the ob gene (directly or indirectly) due to a defective stimulation of the hypothalamic satiety leptin receptor itself, in analogy with the db mouse(19, 20). In four of 20 nonobese children (one girl and three boys) relatively high leptin mRNA expression was observed. Interestingly, in contrast to the others, three of these four children had a hereditary predisposition for moderate obesity. However, they were also the oldest ones in the nonobese group. The elevated leptin mRNA expression in these nonobese children might therefore reflect either a relative “leptin resistance” in prepubertal children, as previously discussed by other authors(13), or reflect a large variation of leptin mRNA expression in nonobese children.